To reflow or re-ball a BGA chip with a dry joint ?

[Fraser]

I know the reflow vs reball discussion has been beaten to death where GPU’s, XBoxes and PS2’s are concerned but I am hoping my question is a little different to those scenarios.

I am very active in the thermal imaging sub forum and it has recently become apparent that some FLIR One G2 Camera dongles have suffered a production defect where small BGA chips have dry joints between the solder ball and the PCB.

I would normally look to remove those IC’s, reball them and refit the original IC to the PCB after it had been cleaned and fluxed. This does take time though. In the case of the FLIR One G2 Camera, they are not really worth the time it would take so I am considering flooding the area under the BGA IC with a good quality no clean flux and carrying out a reflow using a preheater and either Pace Hot Air station or basic Chinese IR rework station. My thought was to get the solder balls flowing and then gently press slightly on the IC to help the dry joints to form a good connection to the PCB.

I can already hear the cries of horror from the knowledgeable techs in this forum  Bear with me though. Has anyone tried this approach to dry joints on BGA chips as opposed to the well known issue with flip chip ‘bumps’ which is very different. We are dealing with a cra**y solder job using lead free solder here. I know re-balling is best but these units are only worn £100 and I can sell the Lepton Core for almost that on its own. I can experiment on one of the 5 units that needs this sort of rework but thought I would check if anyone has experience of reflow like this and whether it is a waste of my time.

This is the thread that has lead to my question.....

https://www.eevblog.com/forum/thermal-imaging/flir-one-gen2-failure-investigation-by-fraser/100/

Post 102 refers.

Thank you in advance

Fraser

[OwO]

If you reflow it you have to add flux. I found that if you apply flux to the 4 sides of the chip, during heating once the flux is liquid the ball structure will wick the flux in. You might need to repeat this multiple times (add flux, heat) until there is sufficient flux under the BGA. Then you turn up the hot air to soldering temperature and do the actual reflow.

[thinkfat]

Don't press on the chip! You'll risk the balls shorting out, and then you're looking into taking the chip off, reballing it and putting it back on. Just ample flux and the right amount of heat should do the job.

[wraper]

Reball only. If there was soldering issue from factory, pad is almost certainly oxidized. Simply melting solder likely won't even restore electrical contact, not to say form a reliable solder joint. Though if BGA is not large, you have steady hands and some skill, then you could remove the chip, tin oxidized pads if there are any and reflow with old solder balls while holding chip with tweezers. Thus saving time on reballing. Certainly exercise not for everyone, but can shorten repair 2-3 times.

[Fraser]

Thanks for the comments everyone.

I have also checked and the BGA chips have epoxy re-enforcement at their edges so that makes life more complicated. I was hoping that a decent flux would remove oxides when preheated to 150C but, as Wrapper states, it is by no means certain. I remain of the belief that these little dongle cameras are more trouble than they are really worth. Not FLIR’s finest moment when they released these less than well built F1G2 units.  At one point in time I was getting batches of 10 passing across my bench for investigation. Some were an easy and quick fix, whilst others were more stubborn, hence I still have 5 awaiting repair. Some were sold as parts purely for the Lepton 3 that resides within.

As ever, I am grateful for anyband all advice gives. You are never too old or experienced to learn from others 

Fraser

[Rasz]

I had some luck with just reflowing bga ram. Make sure your flux wont boil at set temperature.

[Fraser]

Thanks Rasz,

I think I will have a play with one of the dead F1G2 units using just a flux plus reflow method as there is nothing to lose  I could use a reasonably active flux as I have a decent swept frequency ultrasonic tank to clean the flux away afterwards.

I will maybe try the re-balling route on one as well to see if it is worth the effort. I am still being offered dead F1G2 units to repair but have declined all requests to date. That may change if I can repair them reasonably quickly.

Fraser

[EE-digger]

Fraser,

You've been very helpful in the past with my FLIR camera so I thought I'd throw this into the discussion although it's ancient.

If you are going to reflow a package with significant value to it, you may want to consider baking the part to remove moisture content.  This can be based on the original MSL of the part if you can find it, or on general size and thickness.

Baking an existing assembly may not be an option but suspending a controlled hot air source over the part for adequate hours should work.

You can then safely expose it to reflow temps without steam explosions, swelling or other damage.

[Fraser]

EE-digger,

Excellent advice. Thank you 

Fraser